3005/2026/WEB 'Large Language Models (LLMs) and prompt engineering'

The public revelation of ChatGPT in November 2022 and the rapid advancement of large language models (LLMs) have highlighted the potential of this technology to revolutionise investigations and assist law enforcement agencies in addressing the big data challenge.  However, the complexity of this technology has led to its underutilisation by law enforcement, resulting in missed opportunities for large-scale casework in terms of efficacy and efficiency. Through this webinar, law enforcement investigators and analysts gain key knowledge of LLMs and prompt engineering that will empower them to effectively leverage LLMs in their daily operations. In doing so, investigators will be better equipped to implement a variety of techniques to address the challenges they encounter. This webinar is part of a CEPOL learning and training programme which includes other (recorded) webinars and an onsite course.

By the end of this webinar, the audience will be able to:

• Explain the essentials from a large language model in terms of modalities, output, strengths and limitations;
• Detail the differences between small models, medium models and frontier models of LLMs and the conditions how each of them can be used by law enforcement;
• Define engineering of prompts for LLMs and describe the characteristics of good prompts for LLMs;
• Define the term hallucinations related to LLMs and explain why LLMs hallucinate;
• Describe the use cases of LLMs for summarisation, entity extraction, translation, classification, anomaly detection, timeline construction and comparing information.